Testing trophic-functional relationships for modelling farmland diversity and functional dynamics

Lead Research Organisation: Rothamsted Research
Department Name: Agro-Ecology


There is widespread concern that the ongoing decline in UK biodiversity will be exacerbated by further agricultural intensification, such as the adoption of GM technologies. The Defra-funded Farm-scale Evaluations were instigated to address these concerns, and successfully demonstrated that the change in management associated with GMHT crop cultivation had an impact on arable diversity. However, given that arable biodiversity is already in a state of decline, we do not have objective criteria with which to compare the effects of new management practices. Also, as with all empirical studies, it is difficult to predict how the results may vary with different crops, or management systems, or at different spatial and temporal scales. Modelling is therefore required to address these problems, and provide a means of predicting the likely risk of new technology against objectively defined biodiversity criteria. In this proposal we aim to test the hypothesis that the model specification of Hawes et al. (2003) and Bohan et al. (submitted a, b) would lead to generalised functional models that could predict ecosystem dynamics and the impacts of diverse management practice on ecosystem functioning. We use the concept of functional groups to amalgamate weed and invertebrate species into tractable groupings for modelling. These functional group models will take data from the regulatory framework to predict the effects of management on functional diversity at a number of levels of complexity relevant to scientists, the public and policymakers to allow objective decision making on the risks of new farming technologies.

Technical Summary

Over the last 4 decades, catastrophic falls in abundance have been documented in UK farmland for some arable weed, non-pest invertebrate and bird species, driven by the intensification of agricultural management. In the Farmscale Evaluations (FSEs) of GMHT crops a novel technology was exhaustively tested for the risks to the environment (biodiversity), for the first time, prior to introduction. The Advisory Committee on Releases to the Environment (ACRE) recommended that future releases of GM crops, and by inference any new conventional management, 'may need to be scrutinised in terms of their environmental impact'. However, no objective mechanism exists either to trigger such FSE-like experiments or to assess their outcomes. To achieve this it will be necessary to unify the areas of regulation and environmental safety, by bringing together research on management, risk, biodiversity and regulation, using models of farmland ecosystem biodiversity to support decision making. In models of farmland ecosystem biodiversity, some formal level of simplification is necessary. We adopt, for testing, a trophic-functional approach where the crop, weeds and invertebrates are modelled according to their contribution to ecosystem function. This approach is supported by previous modelling and experimental work that indicates it is the value and range of functional attributes represented in a system and not species richness, which determines ecological function and dynamics. The functional group model will consist of individual-based, stochastic, simulation models of the weeds and invertebrates which embody the inherent variability in the translocation of biomass through the food web and consequently the variability in functional diversity through time. The individual based models incorporate stochastic population parameter values, for population processes, that reflect the distribution of population rates across the species in each functional group. By comparison to the weed flora, the highly abundant crop has low variability in population parameters, and will, therefore, be modelled using standard simulation techniques. The functional groups are linked through the assimilation of biomass produced in one functional group by individual consumers in another, limited by primary production. The effects of management are then simply modelled as the changes to the population rate values of the weeds and invertebrates. In this way it will be possible to describe the effects of management on the dynamics of biodiversity, in terms of arable system function. This project will integrate existing data, generate new experimental data where information is missing, define and parameterise new functional group models of arable biodiversity and run simulations to predict the impact of management on arable system function for testing and validation. Specifically, we will: 1. develop a generic model of arable plant and invertebrate functional diversity; 2. predict, for testing, the likely impacts of changes in management practice on arable ecosystem dynamics; 3. validate the model against independently sampled field data for deployment as a predictive decision-making tool.


10 25 50